# Engineered CRISPR/Cas9 enzymes improve discrimination by slowing DNA cleavage to allow release of off-target DNA

Jul 17, 2020

### HypaCas9 and Cas9-HF1 show slow observed rates of cleavage

We begin our kinetic analyses by measuring the enzyme active site concentration for each of the Cas9 variants23,24,25. Measuring the amount of product formed in a titration of enzyme with increasing concentrations of DNA revealed active site concentrations of 31 nM, 26 nM, 23 nM for SpCas9, HypaCas9, and Cas9-HF1, respectively, for enzyme samples with a 100 nM nominal concentration based on absorbance at 280 nm (Supplementary Fig. 1a–d). We also measured the active site concentrations of SpCas9 and HiFiCas925 from Integrated DNA Technologies (IDT) and observed similar concentrations of active enzyme (Supplementary Fig. 1e–f). It is important to note that, in each case, the concentration of DNA required to saturate the signal was equal to the concentration of product formed, which eliminates concerns that some of the enzyme might bind DNA but not react. All subsequent experiments were set up using the concentration of active enzyme determined in the active site titration.

To compare the kinetics of on- or off-target DNA substrates of SpCas9 with the engineered variants, we first examined the time course of target strand (HNH) cleavage for each enzyme (Fig. 1 and Supplementary Fig. 2). Data were fit with either a single- or double-exponential function using Eq. (1) or Eq. (2), respectively.

$$Y = A_1e^{ – lambda _1t} + C$$

(1)

where Y represents concentration of cleavage product, A1 represents the amplitude, and λ1 represents the observed decay rate (eigenvalue)17.

$$Y = A_1e^{ – lambda _1t} + A_2e^{ – lambda _2t} + C$$

(2)

where Y represents concentration of cleavage product, A1 represents the amplitude and λ1 represents the observed rate for the first phase. A2 represents the amplitude and λ2 represents the observed rate for the second phase.

Comparison of the observed cleavage decay rates of on- and off-target substrates by SpCas9 shows that the 3 bp PAM-distal mismatch slows the enzyme 13-fold (from 1 s−1 to 0.076 s−1). Both high-fidelity Cas9 variants dramatically decrease the observed decay rate for cleavage of on-target DNA substrates 21- to 35-fold compared to SpCas9 (0.028 s−1 for HypaCas9 and 0.047 s−1 for Cas9-HF1 vs 1 s−1 for SpCas9). Moreover, HypaCas9 and Cas9-HF1 further reduce the decay rates of off-target DNA cleavage 8- to 290-fold (rates of 0.0033 s−1 and 0.00016 s−1, respectively) relative to their respective rates with on-target DNA. These data demonstrate dramatic changes in the observed rates of cleavage by the engineered variants with on- and off-target DNA. However, these measurements alone do not define changes in specificity, which is a function of the kinetic partitioning of DNA cleavage versus dissociation, encompassing all steps leading up to the first irreversible step in the pathway17, including reversible DNA binding, R-loop formation, HNH domain docking, and DNA cleavage.

### DNA unwinding is largely unchanged with the Cas9 variants

Since our previous work identified R-loop formation as rate-limiting for on-target cleavage and others subsequently suggested that R-loop formation and rewinding rates may dictate enzyme specificity for SpCas9 and Cas9-HF116, we tested whether HypaCas9 would display similar kinetics. To directly measure the rates of R-loop formation for all enzymes, we used a stopped-flow assay based on measuring fluorescence of tCo at −16 nt, a fluorescent tricyclic cytosine analog that is quenched by base stacking in dsDNA so that opening of the duplex results in a large increase in fluorescence. Our control experiments using both our tCo– and 2-AP-labeled base analog substrates at positions −16, −9, and −1 nt, respectively (Supplementary Fig. 3), show that neither of the analogs affect the observed decay rate for DNA cleavage. The chemical structures of tCo and 2AP are much less bulky than larger Cy3 and Cy5 labels and less likely to interfere with enzyme kinetics (Supplementary Fig. 4). In the presence of Mg2+, SpCas9, HypaCas9, and Cas9-HF1 unwind the on-target DNA substrate with nearly identical decay rates (~2 s−1) (Fig. 2). Surprisingly, the decay rate of R-loop formation for off-target DNA substrates for all Cas9 variants was also largely unchanged (between 0.85 s−1 and 2.59 s−1). Therefore, DNA unwinding is not rate-limiting and is not correlated with observed rates of cleavage for the high-fidelity variants, unlike SpCas9.

Given the location of the tCo at −16 nt, it is likely that our measurements reflected a step late in the unwinding process. For comparison, we measured the decay rate for R-loop formation using tCo labeled at a position immediately proximal to the PAM (position −1 nt). After rapidly mixing Cas9-RNA with the on-target DNA substrate, we observed a marked increase in fluorescence as the DNA unwinds the tCo base analog. We fit the data to a double-exponential to determine the major decay rate of 10 s−1 (Fig. 3a–f). Intriguingly, these results indicate that once a PAM site is engaged with the enzyme, initial unwinding of the DNA is faster than observed at −16 nt. These data suggest that the net observed decay rate of unwinding observed at −16 nt may be a function of several fast unwinding step leading to complete unwinding. However, because no lag was observed in the kinetics, it appears that the faster, earlier partial unwinding steps lead to a final rate-limiting step of full unwinding, measured at the −16 nt position. Although further studies are needed to examine the effects of mismatches at earlier stages of unwinding, our current measurement provides the best estimate of the net rate of R-loop formation. The decay rate measured for R-loop formation using this label is indistinguishable for both SpCas9 and the high-fidelity variants on all substrates tested, so the rates for DNA unwinding appear to be unchanged by the engineered Cas9 enzymes.

To correlate the rates of R-loop formation with steps involved in HNH-domain docking onto the target strand, we measured the enzyme conformational change using stopped-flow analysis on a Cas9 which was labeled with Cy3 and Cy5 as described previously18. Briefly, a Cysteine-light version of Cas9 was labeled with Cy3 at amino acid 355 and Cy5 at amino acid 867. The FRET efficiency increases when the HNH domain rearranges to the catalytically active state (Fig. 3g). After rapidly mixing the FRET-pair-labeled Cas9 with a perfectly matched on-target DNA, we observed an increase in FRET efficiency, indicating that the HNH domain was rearranging to a catalytically active state. We measured the decay rate for HNH docking to be ~2.5 s−1 (Fig. 3h), which is similar to single molecule FRET measurements. Surprisingly, the observed rates of R-loop formation (1.5 s−1) and HNH domain docking are very similar, indicating that these steps may be kinetically linked. The slightly faster observed decay rate for HNH domain movement could be due to the reverse reaction as the domain movement comes to equilibrium after or coincident with R-loop formation.

The decay rates of DNA unwinding have also been measured using single molecule methods using a FRET-pair labeled DNA, Cy3 and Cy5 were labeled on position −6 nt of the target strand and -16 nt on the non-target strand, respectively16. Therefore, we also tested the FRET-paired DNA substrates in an attempt to correlate the FRET signal with the observed rates of DNA cleavage. First, we tested the substrate previously used in single molecule studies without the Cy3/Cy5 labels16, which has a two-nucleotide difference with our substrate, specifically, a T substitution at positions −16 and −18. The decay rate of target strand (HNH) cleavage is similar to that of our substrate (0.7 s−1 vs 1 s−1, Supplementary Fig. 5a), so sequence context at this position does not have a significant affect. We also tested cleavage with Cy3/Cy5 labeled DNA with this other sequence, and it showed a similar decay rate as our sequence (0.05 s−1 vs 0.06 s−1, Supplementary Figs. 5b and 6a). Later, we examined the time course of target strand (HNH) cleavage of the Cy3/Cy5 labeled DNA with our sequence for each enzyme (Supplementary Fig. 6). With SpCas9, the reaction of the Cy3/Cy5 labeled DNA follows a single exponential with a markedly reduced decay rate (0.06 s−1 vs 1 s−1) showing a ~17-fold decrease in the observed decay rate for DNA cleavage compared to unlabeled substrate. The engineered HypaCas9 and Cas9-HF1 also exhibited decreased DNA cleavage rates of 0.0046 s−1 and 0.0016 s−1, respectively, which are 5.9-fold and 28.75-fold slower than unlabeled substrates measured under identical conditions. It is clear that interference by the Cy3/Cy5 labels alters the effect of the high-fidelity variants on the DNA cleavage rates. Taken together, labeling of the DNA with bulky Cy3 and Cy5 labels dramatically impacted the enzyme. For these reasons, no further studies were performed using the FRET-pair labeled DNA.

### Cas9 specificity is governed by kinetic partitioning

Since enzyme specificity is a function of all steps leading up to the first largely irreversible step, all events prior to DNA cleavage must be considered17. To measure the intrinsic cleavage rate, we bypassed the normally rate-limiting R-loop formation step by preincubating SpCas9 with off-target DNA in the absence of Mg2+ to allow binding and conformational changes to come to equilibrium to form an SpCas9.DNA complex. We then initiated the chemical reaction by the addition of 10 mM Mg2+. In our previous studies, R-loop formation was rate-limiting with SpCas9 reacting with on-target DNA, and the intrinsic cleavage decay rate was faster when measured using the preincubation protocol. Here, with off-target DNA, the rates of HNH and RuvC cleavage were measured to be 0.12 s−1 and 0.14 s−1, respectively (Supplementary Fig. 7c, d, and Supplementary Fig. 8), which are much slower than the rates of R-loop formation. Intriguingly, these results show that the rate-limiting step in the enzyme pathway of SpCas9 with an off-target is DNA cleavage since the decay rate for R-loop formation we measured was 0.85 s−1 (Fig. 2b). These data indicate that discrimination is based, at least in part, on a change in the identity of the rate-limiting step in comparing on- and off-target DNA.

We repeated these experiments with HypaCas9 and Cas9-HF1 with on- or off-target DNA. These results define observed rates for HNH cleavage of 0.035 s−1 and 0.0023 s−1 for HypaCas9 with on- and off-target substrates, respectively (Supplementary Fig. 7g, k). Observed cleavage rates of Cas9-HF1 for on- and off-target substrates were measured as 0.038 s−1 and 0.00014 s−1, respectively (Supplementary Fig. 7o, q). These observed cleavage rates are somewhat faster than those measured with the simultaneous addition of DNA and Mg2+, indicating that some step other than R-loop formation but preceding DNA cleavage may slow the observed decay rate. Nonetheless, these results show that the observed cleavage rates for on-target DNA are reduced ~100-fold for both HypaCas9 and Cas9-HF1 relative to SpCas9. For off-target DNA, the observed cleavage rates are reduced 50- or 860-fold for HypaCas9 or Cas9-HF1, respectively, relative to SpCas9.

Discrimination is not defined solely by the relative rates of DNA cleavage. Rather, because R-loop formation is fast, discrimination is a function of the kinetic partitioning between the rates of DNA release versus cleavage17,26,27. In order to quantify the kinetic partitioning, we incubated enzyme and labeled DNA in the absence of Mg2+, which allows for R-loop formation to come to equilibrium, but prevents catalysis15. We then added Mg2+ to initiate the reaction in the presence of a large excess of identical, unlabeled on-target DNA to serve as a trap. Comparison between parallel experiments performed in the presence and absence of the DNA trap provides an estimate for the fractional kinetic partitioning for dissociation versus cleavage of bound DNA (Fig. 4a). Once SpCas9 was bound to on-target DNA, it was cleaved rapidly, and the DNA trap had little effect (Fig. 4b). In contrast, 33% of the off-target DNA disassociated from the enzyme, while ~67% of the DNA was committed to cleavage (Fig. 4c and Supplementary Fig. 9a). These results show that SpCas9 discriminates against the PAM-distal mismatched DNA by decreasing the rate of cleavage, which increases the fraction of DNA that is released rather than cleaved. However, the effect is small so the dissociation rate appears to be slower than cleavage.

Next, we examined the kinetic partitioning for HypaCas9 and Cas9-HF1 bound to on-target DNA (Fig. 4d, f, Supplementary Fig. 9b, d). Our results with HypaCas9 and Cas9-HF1 show that ~75% and ~92% of the on-target DNA was cleaved in the presence of the trap, respectively. The percentage cleaved is smaller than for wild-type SpCas9 because the observed intrinsic cleavage decay rate for on-target DNA by HypaCas9 (0.035 s−1) and Cas9-HF1 (0.038 s−1) was 100-fold slower than with SpCas9 (4.3 s−1). This slower cleavage rate gives time for a small fraction (8 to 25%) of the on-target DNA to dissociate before it is cleaved.

Kinetic partitioning to favor dissociation was enhanced when HypaCas9 and Cas9-HF1 react with off-target DNA because the cleavage rates were further reduced to 0.0023 s−1 and 0.00014 s−1, respectively (Fig. 4e, g, Supplementary Fig. 9c, e). These rates are 50- to 860-fold slower, respectively, compared to SpCas9 on an off-target substrate. Accordingly, only ~24% and ~10% of the bound off-target DNA was committed to going forward for cleavage by HypaCas9 and Cas9-HF1, respectively, in the presence of trap DNA. Taken together, these results show that the engineered high-fidelity variants acquired improved specificity against the PAM-distal mismatched DNA through a markedly decreased rate of cleavage, which alters kinetic partitioning to favor release rather than cleavage of the bound substrate for both on- and off-target DNA, but the effect is larger for off-target DNA. Calculation of the apparent dissociation rate constants using Eq. (3) shows that the high-fidelity variants do not increase the apparent rate of DNA release (Supplementary Table 1).

$${mathrm{Cleavage}},{mathrm{probability}} = frac{{k_{chem}}}{{k_{off} + k_{chem}}}$$

(3)

Rather, the increased discrimination is entirely attributed to decreases in the apparent rate constant for cleavage.

In our descriptions of the kinetics of Cas9 enzymes thus far, the observed decay rates of reactions were estimated based on fitting data to exponential functions, which yield eigenvalues that are usually complex functions of multiple rate constants17. To fully understand the kinetics and mechanism, all of the experiments for each enzyme and DNA substrate were fit globally based on numerical integration of the rate equations using a single unified model (Fig. 5 and Supplementary Figs. 1014). Global data fitting shows that our minimal model (Fig. 5f, inset) accounts for our data and each of the rate constants is well defined based on confidence contour analysis testing the extent to which each parameter is constrained by the data (Fig. 6)17,28. Note in particular that our model accounts for the biphasic traces seen in some experiments without having to invoke heterogeneity, and it provides intrinsic rate constants for each step in the pathway including R-loop formation and HNH and RuvC cleavage events. Although it is likely that additional, unresolved structural rearrangements may be required for alignment of catalytic residues for DNA cleavage, these have not yet been resolved as kinetically distinct events. The current model represents a minimal pathway necessary and sufficient to account for all available data, and can be easily expanded as new information becomes available to define additional steps in the pathway.

To understand enzyme specificity, rate constants must be interpreted in the context of all kinetically relevant steps as illustrated in the free energy profile (calculated using Eq. (4) from rate constants in Table 1).

$$Delta {mathrm{G}}^ddagger = RTleft( {{mathrm{ln}}left( {{mathrm{A}} ast kT/h} right)-{mathrm{ln}}left( {k_{obs}} right)} right),{mathrm{kcal}}/{mathrm{mol}}$$

(4)

The transmission coefficient A = 0.001

Because global data fitting provides the rate constants for each relevant step (Figs. 5 and 6), we can construct a bona fide free energy profile (Fig. 7b, and Supplementary Figs. 1014). The free energy profiles comparing SpCas9, HypaCas9, and Cas9-HF1 show a change in rate-limiting and specificity-determining steps. Enzyme specificity is defined by kcat/Km and is given by the highest overall barrier relative to the starting state, while the maximum rate, kcat, is defined by the highest local barrier relative to the preceding state, attenuated by any preceding rapid-equilibrium steps. Because the rate constants for R-loop formation do not change significantly with different substrates and enzyme variants, specificity is governed largely by the kinetic partitioning of the Cas9 R-loop, EDH state in our kinetic model (Fig. 5f, inset), to either go forward resulting in irreversible cleavage versus release via re-annealing of the DNA and ejection from the enzyme. The higher overall barriers for cleavage seen with the high-fidelity variants increase the kinetic partitioning probability to favor dissociation of the DNA (Fig. 6c).